Author Topic: Fluke 8500A 6.5 digit problems/modifications and repair  (Read 2711 times)

0 Members and 1 Guest are viewing this topic.

Offline Enigma-manTopic starter

  • Regular Contributor
  • *
  • Posts: 73
  • Country: ca
Fluke 8500A 6.5 digit problems/modifications and repair
« on: August 20, 2016, 04:31:16 am »
After looking at many different threads here many people enjoy the efforts of those doing repairs on various types of equipment. 
I would like to share my Fluke 8500A story. In the last week of July, 2015 I was given a Fluke 8500A 6.5 digit multimeter, for free.
A friend had two. A newer 8502A with all the plugin cards and a volts only 8500A. I was happy to adopt the volts only meter. 
It was in remarkable shape as it spent its entire life in service with NAV Canada. It had age problems, the common yellowing of the ivory
plastic that plagues all older equipment. No big deal, I fixed that cosmetic issue a few days later. It seemed to work okay but I noticed it was
very warm on the top left side of the unit after it was left on for a few hours.  There is no ventilation whatsoever.  After removing
the top cover I felt around the heatsinks and one was so hot you could barely touch it.  Not good... I also noticed an oblong bridge
rectifier near the back was rather hot and could see the pc board was brown as a result of the heat after many years of service. 
I was about to turn off the Fluke and noticed the power switch was sticking. Another minor problem.  My 7704A Tektronix scope was on
waiting for something to do.  I had some unused miniature four pin 80.0000Mhz xtal oscillators that I was going to check. When I applied
power to my protoboard with one of the oscillators ready to test,the Fluke went nuts.  Erratic numbers all over the place.  I shut off
the 80MHz oscillator and all is fine.  I had an idea as to the problem and removed the power supply module by sliding it out.  Turning
the power supply over revealed a very crappy ac line filter. Two small toroids and two ceramic caps.  Forty years ago, computer equip-
ment was running at 1, 2 or 4MHz.  No need for elaborate filtering so what was present was good enough back then, not now.
Attached pictures show what was done along with some text boxes for explanation.  I won't repeat what's in the pictures.
The power supply issues were solved and it ran much cooler, but there were issues elsewhere as I would find out a few weeks later.
The 2708's needed three voltages, +5, +12 and -5V to operate.  The eprom board had a 2708 that was running hot. There was no room
for a heatsink on the board.  The cpu board and eprom board are connected together and inside a plastic box as are all plug in
boards. All that could dissipate heat was what you see in the picture. I removed the 7905 and attached it to a spot on the aluminum
chassis.
I left the 8500A running for several days at a time, measuring 10.000V from an AD688 precision 10.000V reference.  One day the
display blanked and hitting reset did nothing. Powering off then on again got the unit working again.  This was to be a random
occurrence over the past year. I didn't like the 2708's.  They generated a lot of heat and eventually one or more would give out
so I decided to reprogram their contents to eight 2732's.  This idea turned out to be more than I bargained for.  I couldn't find
an eprom programmer to do the job.  I had an old Apparat programmer from my Model 3/4 days. It could read and write 2708's but...
I didn't have a working TRSDOS, Multidos, or any DOS that would boot from a Model 3 or 4. About to abandon the idea I was when I
remembered an old CoCo1 was made into a dedicated eprom programmer.  It needed an old TV or a composite video monitor to work.
A Panasonic 9 inch B/W monitor was still around and I powered up the CoCo1.  It still worked.  At some point way back when I
had enough brains to have written "53248" on the pc board (in pencil).  I typed "exec 53248" and hit enter. Up came the menu.
I was in business.  It took longer to erase a bunch of 2732's pulled from who knows what and find out some were no good.  I would
read each 2708 one at a time and punch out the contents to a 2732.  A few hours later they were all done and the 8500A was up and
running.  The blanking/dead mode kept coming back and I would change out suspect IC's according to the troubleshooting section.

The images contain the info for what was done.

To be continued very soon...
 
The following users thanked this post: Storm

Offline Enigma-manTopic starter

  • Regular Contributor
  • *
  • Posts: 73
  • Country: ca
Re: Fluke 8500A 6.5 digit problems/modifications and repair
« Reply #1 on: August 20, 2016, 05:34:03 pm »
Part 2:

A problem was noticed when I changed out the first IC.  There was something wrong with the board itself. Take a look at the second
pic. After desoldering, there was no plating on the pads, just discoloured copper pads that won't take solder.  The plating or tinning
was no good on the upper side of the board.  The entire component side is like this.  I changed out I had to use an eraser made for ink to scour the copper enough to take solder.  The problem
(not confirmed yet but I'm pretty sure) is likely the way the boards are mated together.  There are two pin strips with square pins on the
eprom board that mate with two connectors on the cpu board. The second pic also shows the eprom board connector.
I believe the problem is one or both connectors on the cpu board (now removed) have sprung contacts causing intermittent connection(s)
and loss of control.  There was no socket at U22. I desoldered the area to show what the whole board looks like from the component side.
I will again scour the pads with the ink eraser, apply solder to the areas then use my Pace desoldering station to suck up the
solder leaving a tinned area for the 24 pin machine socket.  I have a 24 pin dip header that will have its wires soldered to the eprom board.
The strip rows on the eprom board have been removed. As you can see in the third pic, the eprom board doesn't have the same problem as the
cpu board. There will be a secure mechanical connection between the dip header and machine pin socket.

I have seen hundreds of double sided boards but never one with a defect such as this.  I wonder how many 8500's with controller problems
have this same issue.  I doubt I am the only one with a pc board like this.

The white IC is an 8080A cpu  clocked at a whopping 1.700 MHz  ;D

Anyway, I am going to wire up the dip header and go from there.

 
« Last Edit: August 20, 2016, 05:39:01 pm by Enigma-man »
 
The following users thanked this post: Storm

Offline bitseeker

  • Super Contributor
  • ***
  • Posts: 9057
  • Country: us
  • Lots of engineer-tweakable parts inside!
Re: Fluke 8500A 6.5 digit problems/modifications and repair
« Reply #2 on: August 20, 2016, 10:04:14 pm »
That's quite a project, E-man. Keep up the good work. Looking forward to seeing it up and running as well as the restored plastic enclosure.

It was funny to hear that a CoCo1 saved the day. My first computer was a CoCo1. I still have it, but it just outputs garbage characters on the screen. That's a project for another day.
TEA is the way. | TEA Time channel
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf